Thread: Backend crash
The following sequence causes PostgreSQL to crash on my Linux box: -- Any example table here will do CREATE TABLE numbers ( value int ); INSERT INTO numbers VALUES(1); INSERT INTO numbers VALUES(2); INSERT INTO numbers VALUES(3); INSERT INTO numbers VALUES(4); -- Now the crash sequence BEGIN; SELECT * INTO TEMP numbers2 FROM numbers; SELECT * FROM numbers2; -- Bang!!! -- Can anybody reproduce this? Thanks, Mark
Mark Volpe <volpe.mark@epamail.epa.gov> writes: > BEGIN; > SELECT * INTO TEMP numbers2 FROM numbers; > SELECT * FROM numbers2; > -- Bang!!! -- > Can anybody reproduce this? Yah :-(. There are several nasty bugs in 6.5.*'s support for TEMP tables. We believe they are all fixed for 7.0, but the changes are intertwined with a lot of other stuff, and so back-patching the fixes didn't look very practical. The only good news is that 7.0 goes beta next week... regards, tom lane
> The following sequence causes PostgreSQL to > crash on my Linux box: > > -- Any example table here will do > > CREATE TABLE numbers ( value int ); > INSERT INTO numbers VALUES(1); > INSERT INTO numbers VALUES(2); > INSERT INTO numbers VALUES(3); > INSERT INTO numbers VALUES(4); > > -- Now the crash sequence > > BEGIN; > SELECT * INTO TEMP numbers2 FROM numbers; > SELECT * FROM numbers2; > > -- Bang!!! -- > > Can anybody reproduce this? Fixed in current source tree. We had some problems with temp tables created inside transactions. Next release will fix this. -- Bruce Momjian | http://www.op.net/~candle pgman@candle.pha.pa.us | (610) 853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania19026
Thanks, guys! SELECT * INTO <good-old-unique-name-based-on-process-id> ... works just fine for me :-) Tom Lane wrote: > > Mark Volpe <volpe.mark@epamail.epa.gov> writes: > > BEGIN; > > SELECT * INTO TEMP numbers2 FROM numbers; > > SELECT * FROM numbers2; > > -- Bang!!! -- > > Can anybody reproduce this? > > Yah :-(. There are several nasty bugs in 6.5.*'s support for TEMP > tables. We believe they are all fixed for 7.0, but the changes are > intertwined with a lot of other stuff, and so back-patching the fixes > didn't look very practical. > > The only good news is that 7.0 goes beta next week... > > regards, tom lane > > ************
> Thanks, guys! > > SELECT * INTO <good-old-unique-name-based-on-process-id> ... > works just fine for me :-) There is a section in my book about it. > > Tom Lane wrote: > > > > Mark Volpe <volpe.mark@epamail.epa.gov> writes: > > > BEGIN; > > > SELECT * INTO TEMP numbers2 FROM numbers; > > > SELECT * FROM numbers2; > > > -- Bang!!! -- > > > Can anybody reproduce this? > > > > Yah :-(. There are several nasty bugs in 6.5.*'s support for TEMP > > tables. We believe they are all fixed for 7.0, but the changes are > > intertwined with a lot of other stuff, and so back-patching the fixes > > didn't look very practical. > > > > The only good news is that 7.0 goes beta next week... > > > > regards, tom lane > > > > ************ > > ************ > > -- Bruce Momjian | http://www.op.net/~candle pgman@candle.pha.pa.us | (610) 853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania19026